This Repo. contains code used for data collection, cleaning as well as model codes for running CNNs on Landsat Images at village level
Description of Files:
decidingDimensionsAndClusteringVillages.ipynb: This is used to run the base model and trying to decide Village Input Dimensions of CNN.
downloadLandsat.js: This is run on Google Earth Engine to extract images of states as a whole. It automatically gets split during downloading.
download.py: This was used to quickly download the extracted images on Google Drive to Azure VM very quickly.
extractTiff.py: This is used to clip the villages from the state images. It uses raster.io package.
inceptionPretrain.py: This is used to transfer learn followed by fine-tuning for 2 epochs each.
cnnMfromScratch.py: This is used to train a neural network from scratch using CNN-M architechture with BN added after each layer.
ayushgupt/landsatCNN
This Repo. contains code used for data collection and cleaning as well as model codes for running CNNs on Landsat Images at village level
Jupyter Notebook